Abstract—Internet Of Things (IoT) is the term which has been popular nowdays as an increasing number of users. Statistically results shows that in the future it will increasing more and more. But because of this large number of uses of users, to maintain their high degree of security is something that is critical. In this research we has been liked to improve the security of IOT devices through applying various Machine Learning Algorithms and through some efficient engineering techniques. In this Paper, we have set up an approach to detect botnet of IOT devices using three ML Algorithms that are: Support vector machine second one is Naïve Bayes and Third one is Decision Tree which all are Supervised Learning Algorithms. And for the purpose to detect Bot we have been using different Bot Datasets. After a number of pre-processing steps, we feed the pre-processed data to our supervised algorithms that can achieve a good precision score that is approximately 77–99%. The SVM achieves the best accuracy score, approximately 99% in every dataset, and Naïve Bayes accuracy score varies from 91% to 98%; however, the Decision Tree achieves lowest accuracy score that is from 77% to 99%. Our algorithms are cost-effective and provide good accuracy in short execution time.
Keywords— IoT, IoT botnet, Botmster, Bot attack, IoT devices, P2P, Datasets, DDoS, Feature Extraction, ML algorithms, Cyber Security, malicious, data pre-processing.
|
DOI:
10.17148/IJARCCE.2023.125208
